Zulip Chat Archive

Stream: maths

Topic: Ideal is intersection of comap_map along localisations


Geno Racklin Asher (Jun 16 2024 at 11:30):

I'm currently attempting to formalise Prop. II.3.2 from Hartshorne's Algebraic Geometry, which states that a scheme is locally noetherian iff it has a cover by noetherian affines. I've done all the scheme-theoretic parts of the proof and have reduced it to proving the following algebraic statement (as in Hartshorne's proof):

variable {R : Type u} [CommRing R] (S : Finset R) (hS: Ideal.span (α := R) S = )

def mcm (I : Ideal R) (s : R) :=
  let f := algebraMap R (Localization.Away s)
  Ideal.comap f (Ideal.map f I)

lemma ideal_eq_sect (I : Ideal R) : I.carrier =  (s : S), mcm I s := by
  let floc s := algebraMap R (Localization.Away (M := R) s)
  apply Set.ext
  intro x
  apply Iff.intro
  . intro hx
    apply Set.mem_iInter.mpr
    intro s
    exact Ideal.le_comap_map hx
  . intro hx
    have hx:  s : S, x  mcm I s := by
      intro s
      apply Set.mem_iInter.mp hx s
    have :  s : S,  n : , (floc s x) * (floc s s ^ n)  Ideal.map (floc s) I := by
      intro s
      have hx := hx s
      sorry
    sorry

But i'm now unsure how to proceed in formalising the remainder of the lemma. Hartshorne's proof is as follows:

image.png

Any advice for formalising this line of argument would be much appreciated!

Yaël Dillies (Jun 16 2024 at 11:32):

Hey Geno! :wave:

Andrew Yang (Jun 16 2024 at 11:37):

docs#IsLocalization.mem_map_algebraMap_iff is useful here.

Andrew Yang (Jun 16 2024 at 11:56):

And also docs#Submodule.mem_of_span_eq_top_of_smul_pow_mem

Andrew Yang (Jun 16 2024 at 12:01):

Those two should take you to the goal, but if you want to see my approach:

My approach

Geno Racklin Asher (Jun 16 2024 at 15:21):

Thanks!

Kevin Buzzard (Jun 17 2024 at 12:49):

I hope you'll be PRing the result @Geno Racklin Asher !

Geno Racklin Asher (Jun 17 2024 at 13:01):

Yes, my plan is to try and formalise a couple more basic results from Stacks, namely that (locally) noetherian schemes are quasi-separated and have (locally) noetherian underlying topological spaces; but yes it would be great to get as much of that as possible PR'd into mathlib!

Geno Racklin Asher (Jun 19 2024 at 21:32):

Now PR'd at #13972!


Last updated: May 02 2025 at 03:31 UTC